1. Field of the Invention
The present invention is related to a detecting method in a non-volatile memory, and more particularly, to a detecting method capable of detecting the disturb phenomena between neighboring blocks in the non-volatile memory.
2. Description of the Prior Art
Flash memory is a non-volatile memory which can perform numerous times of read and write functions. Meanwhile, as the data capacity of the flash memory increased, and the data access speed advanced, applications of the flash memory have been getting even more popular; from small-sized products, like memory card, keychain drive to larger-sized gadgets, like mobile phone, video game machine, digital camera and many other electronic devices, the flash memory almost exists in all kinds of product as a data storage device.
The flash memory comprises numerous cells, each having a specially designed floating gate, and according to principles of quantum mechanics, electronic charges can be transported into the floating gate by an effect known as tunnel injection, and the electronic charges can also be taken out of the floating gate by a reverse effect named tunnel release. The tunnel injection and the tunnel release jointly constitute the basic mechanism to store or erase data in the flash memory. Although the flash memory has many advantages in price and performance, according to its operating principles, the flash memory has a major limitation that it can only be accessed in blocks, which is similar to the data access method to a hard disk drive. On the other hand, the flash memory device might have manufacturing faults resulted from imperfections existing in the memory manufacturing process, such that data being written is inconsistent of data being read from the memory. Moreover, peculiar things could happen between neighboring blocks in the flash memory: data stored in a memory block can be unexpectedly altered by a neighboring block, while erase action(s) is applied to the neighboring block. This phenomena is generally referred as disturb (or crosstalk) phenomena of a flash memory.
Generally speaking, the detection method used for detecting manufacturing faults of a flash memory is for checking the data validity of a block itself, and not for checking the disturb phenomena between neighboring blocks. Basically, according to the prior art, the testing method generally comprises erasing memory blocks, writing data into memory blocks, reading the data out to check its validity by comparing the data being read, and then marking the failed block if the data is inconsistent. As can be understood, this kind of testing method is not designed for checking the disturb-type of manufacturing faults in the flash memory. Please refer to FIG. 1, which illustrates a schematic diagram of a flash memory 10. The flash memory 10 comprises Block_0˜Block_2N of memory blocks, where a disturb phenomena exists between Block_K and Block_K+1. In other words, while the erase operation is applied to Block_K, a portion of data stored in Block_K+1 could be changed anomalously owing to the disturb phenomena. Meanwhile, if the erase operation is applied to Block_K+1, then data stored in Block_K could also be changed. Simply speaking, between neighboring blocks, say Block_K and Block_K+1, if a selected memory block is being erased, the data of the other neighboring block can be altered unwillingly through an electromagnetic action imposed by the block being erased. However, if the testing method of the prior art is adopted, the data in Block_K will not be changed by erasing and writing test data to Block_K itself, and the data in Block_K+1 will not be changed by erasing and writing test data to Block_K+1, either. The testing result will indicate Block_K and Block_K+1 are both functional (which are incorrect), and the disturb phenomena between Block_K and Block_K+1 cannot be effectively detected according to the prior art. In other words, the testing method in the prior art is improper to be used to detect the disturb-type manufacturing faults of the flash memory 10.